Event Reconstruction in a Liquid Xenon Time Projection Chamber with an Optically-Open Field Cage
T. Stiegler, S. Sangiorgio, J.P. Brodsky, M. Heffner, S. Al Kharusi, G. Anton, I.J. Arnquist, I. Badhrees, P.S. Barbeau, D. Beck, V. Belov, T. Bhatta, A. Bolotnikov, P.A. Breur, E. Brown, T. Brunner, E. Caden, G.F. Cao, L. Cao, C. Chambers, B. Chana, S.A. Charlebois, M. Chiu, B. Cleveland, M. Coon, A. Craycraft, J. Dalmasson, T. Daniels, L. Darroch, A. De St. Croix, A. Der Mesrobian-Kabakian, K. Deslandes, R. DeVoe, M.L. Di Vacri, J. Dilling, Y.Y. Ding, M.J. Dolinski, A. Dragone, J. Echevers, F. Edaltafar, M. Elbeltagi, L. Fabris, D. Fairbank, W. Fairbank, J. Farine, S. Ferrara, S. Feyzbakhsh, G. Gallina, P. Gautam, G. Giacomini, D. Goeldi, R. Gornea, G. Gratta, E.V. Hansen, E.W. Hoppe, J. Hößl, A. House, M. Hughes, A. Iverson, A. Jamil, M.J. Jewell, X.S. Jiang, A. Karelin, L.J. Kaufman, T. Koffas, R. Krücken, A. Kuchenkov, K.S. Kumar, Y. Lan, A. Larson, K.G. Leach, B.G. Lenardo, D.S. Leonard, G. Li, S. Li, Z. Li, C. Licciardi, P. Lv, R. MacLellan, N. Massacret, T. McElroy, M. Medina-Peregrina, T. Michel, B. Mong, D.C. Moore, K. Murray, P. Nakarmi, C.R. Natzke, R.J. Newby, K. Ni, Z. Ning, O. Njoya, F. Nolet, O. Nusair, K. Odgers, A. Odian, M. Oriunno, J.L. Orrell, G.S. Ortega, I. Ostrovskiy, et al. (51 additional authors not shown)
EEvent Reconstruction in a Liquid Xenon Time Projection Chamber with anOptically-Open Field Cage
T. Stiegler a,1 , S. Sangiorgio a , J.P. Brodsky a , M. He ff ner a , S. Al Kharusi b , G. Anton c , I.J. Arnquist d , I. Badhrees e,2 ,P.S. Barbeau f , D. Beck g , V. Belov h , T. Bhatta i , A. Bolotnikov j , P.A. Breur k , E. Brown l , T. Brunner b,m , E. Caden n,3 ,G.F. Cao o,4 , L. Cao p , C. Chambers b , B. Chana e , S.A. Charlebois q , M. Chiu j , B. Cleveland n,3 , M. Coon g ,A. Craycraft r,5 , J. Dalmasson s , T. Daniels t , L. Darroch b , A. De St. Croix u,m , A. Der Mesrobian-Kabakian n ,K. Deslandes q , R. DeVoe s , M.L. Di Vacri d , J. Dilling m,u , Y.Y. Ding o , M.J. Dolinski v , A. Dragone k , J. Echevers g ,F. Edaltafar m , M. Elbeltagi e , L. Fabris w , D. Fairbank r , W. Fairbank r , J. Farine n , S. Ferrara d , S. Feyzbakhsh x ,G. Gallina u,m , P. Gautam v , G. Giacomini j , D. Goeldi e , R. Gornea e,m , G. Gratta s , E.V. Hansen v , E.W. Hoppe d ,J. H¨oßl c , A. House a , M. Hughes y , A. Iverson r , A. Jamil z , M.J. Jewell s,6 , X.S. Jiang o , A. Karelin h , L.J. Kaufman k ,T. Ko ff as e , R. Kr¨ucken u,m , A. Kuchenkov h , K.S. Kumar x , Y. Lan u,m , A. Larson aa , K.G. Leach ab , B.G. Lenardo s ,D.S. Leonard ac , G. Li o , S. Li g , Z. Li z , C. Licciardi n , P. Lv o , R. MacLellan i , N. Massacret m , T. McElroy b ,M. Medina-Peregrina b , T. Michel c , B. Mong k , D.C. Moore z , K. Murray b , P. Nakarmi y , C.R. Natzke ab , R.J. Newby w ,K. Ni ad , Z. Ning o , O. Njoya ae , F. Nolet q , O. Nusair y , K. Odgers l , A. Odian k , M. Oriunno k , J.L. Orrell d , G.S. Ortega d ,I. Ostrovskiy y , C.T. Overman d , S. Parent q , A. Piepke y , A. Pocar x , J.-F. Pratte q , V. Radeka j , E. Raguzin j , S. Rescia j ,F. Reti`ere m , M. Richman v , A. Robinson n , T. Rossignol q , P.C. Rowson k , N. Roy q , R. Saldanha d , K. Skarpaas VIII k ,A.K. Soma v , G. St-Hilaire q , V. Stekhanov h , X.L. Sun o , M. Tarka x , S. Thibado x , A. Tidball l , J. Todd r , T.I. Totev b ,R. Tsang y , T. Tsang j , F. Vachon q , V. Veeraraghavan y , S. Viel e , G. Visser af , C. Vivo-Vilches e , J.-L. Vuilleumier ag ,M. Wagenpfeil c , T. Wager r , M. Walent n , Q. Wang p , W. Wei o , L.J. Wen o , U. Wichoski n , M. Worcester j , S.X. Wu s ,W.H. Wu o , X. Wu p , Q. Xia z , H. Yang p , L. Yang ad , O. Zeldovich h , J. Zhao o , Y. Zhou p , T. Ziegler c a Lawrence Livermore National Laboratory, Livermore, CA 94550, USA b Physics Department, McGill University, Montr´eal, Qu´ebec H3A 2T8, Canada c Erlangen Centre for Astroparticle Physics (ECAP), Friedrich-Alexander University Erlangen-N¨urnberg, Erlangen 91058, Germany d Pacific Northwest National Laboratory, Richland, WA 99352, USA e Department of Physics, Carleton University, Ottawa, Ontario K1S 5B6, Canada f Department of Physics, Duke University, and Triangle Universities Nuclear Laboratory (TUNL), Durham, NC 27708, USA g Physics Department, University of Illinois, Urbana-Champaign, IL 61801, USA h Institute for Theoretical and Experimental Physics named by A. I. Alikhanov of National Research Center “Kurchatov Institute”, Moscow117218, Russia i Department of Physics and Astronomy, University of Kentucky, Lexington, Kentucky 40506, USA j Brookhaven National Laboratory, Upton, NY 11973, USA k SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA l Department of Physics, Applied Physics and Astronomy, Rensselaer Polytechnic Institute, Troy, NY 12180, USA m TRIUMF, Vancouver, British Columbia V6T 2A3, Canada n Department of Physics, Laurentian University, Sudbury, Ontario P3E 2C6 Canada o Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China p Institute of Microelectronics, Chinese Academy of Sciences, Beijing 100029, China q Universit´e de Sherbrooke, Sherbrooke, Qu´ebec J1K 2R1, Canada r Physics Department, Colorado State University, Fort Collins, CO 80523, USA s Physics Department, Stanford University, Stanford, CA 94305, USA t Department of Physics and Physical Oceanography, University of North Carolina at Wilmington, Wilmington, NC 28403, USA u Department of Physics and Astronomy, University of British Columbia, Vancouver, British Columbia V6T 1Z1, Canada v Department of Physics, Drexel University, Philadelphia, PA 19104, USA w Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA x Amherst Center for Fundamental Interactions and Physics Department, University of Massachusetts, Amherst, MA 01003, USA y Department of Physics and Astronomy, University of Alabama, Tuscaloosa, AL 35487, USA z Wright Laboratory, Department of Physics, Yale University, New Haven, CT 06511, USA Corresponding author [email protected] also at King Abdulaziz City for Science and Technology, Riyadh, Saudi Arabia also at SNOLAB, Ontario, Canada also at University of Chinese Academy of Sciences, Beijing, China now at Intel, Portland, OR, USA now at Yale University, New Haven, CT, USA Preprint submitted to Elsevier September 23, 2020 a r X i v : . [ phy s i c s . i n s - d e t ] S e p a Department of Physics, University of South Dakota, Vermillion, SD 57069, USA ab Department of Physics, Colorado School of Mines, Golden, CO 80401, USA ac IBS Center for Underground Physics, Daejeon 34126, Korea ad Physics Department, University of California, San Diego, La Jolla, CA 92093, USA ae Department of Physics and Astronomy, Stony Brook University, SUNY, Stony Brook, NY 11794, USA af Department of Physics and CEEM, Indiana University, Bloomington, IN 47405, USA ag LHEP, Albert Einstein Center, University of Bern, Bern CH-3012, Switzerland
Abstract nEXO is a proposed tonne-scale neutrinoless double beta decay (0 νββ ) experiment using liquid
Xe (LXe) in aTime Projection Chamber (TPC) to read out ionization and scintillation signals. Between the field cage and the LXevessel, a layer of LXe (“skin” LXe) is present, where no ionization signal is collected. Only scintillation photons aredetected, owing to the lack of optical barrier around the field cage. In this work, we show that the light originatingin the skin LXe region can be used to improve background discrimination by 5% over previous published estimates.This improvement comes from two elements. First, a fraction of the γ -ray background is removed by identifying lightfrom interactions with an energy deposition in the skin LXe. Second, background from Rn dissolved in the skinLXe can be e ffi ciently rejected by tagging the α decay in the Bi-
Po chain in the skin LXe.
Keywords:
Neutrinoless double beta decay, Liquid xenon detectors, Time-projection chambers, Monte Carlomethods
1. Introduction
Neutrinoless double beta decay (0 νββ ) is a second-order weak transition that is predicted to occur in severaleven-even nuclei [1] if neutrinos are Majorana fermions.The observation of this process would indicate leptonnumber violation, demonstrate the Majorana nature ofneutrinos [2], and provide valuable information aboutthe absolute scale of the neutrino mass spectrum.nEXO is a proposed tonne-scale detector that will use ∼ νββ in Xe. A detailed sensitivity analysis [3]has been published along with a pre-conceptual designof nEXO [4]. nEXO measures ionization and scintil-lation signals [5] from the LXe volume inside a fieldcage of evenly-spaced field-shaping rings that establishthe required TPC drift electric field. The scintillationlight is collected by silicon photomultipliers (SiPMs)on the cylindrical “barrel of the detector and the ioniza-tion electrons by charge collection tiles on the anode.The field-cage assembly is surrounded by an insulatinglayer of LXe, referred to as the skin LXe, where no ion-ization signal can be collected. A scintillation signal,on the other hand, is collected from interactions bothinside and outside the field cage. In nEXO, approxi-mately 25% of the liquid xenon is located in the skinLXe region. The question addressed in this paper is toidentify the best way to handle the scintillation light col- lected from the skin LXe, and to demonstrate that it canbe used to improve background discrimination.The field cage surrounded by an insulating skin ofLXe is a typical design feature in detectors of this type,but di ff erent approaches have been taken with regard tolight collection in the skin LXe. One approach has pho-todetectors installed on the ends of the cylindrical vol-ume and a way to optically isolate the LXe volume in-side the field cage from the skin LXe outside [6–8]. Werefer to this arrangement, with two distinct optical vol-umes, as a closed-field-cage design. nEXO dispenseswith the optical barrier around the TPC field cage, and istherefore an ”open-field-cage” design with a single op-tical volume. Fig. 1 shows a cross-section of the nEXOdetector.By dispensing with the optical shield, nEXO removesa potential source of radioactive and chemical contami-nation and improves LXe circulation. In this paper, weshow that the absence of a light barrier around the fieldcage does not prevent the use of light originating in theskin LXe region to classify events. It can be used to im-prove background discrimination by 5% over previousestimates [3].The previous estimates did not evaluate the impli-cations of the open-cage configuration but rather madetwo assumptions. First, it was assumed that α particlesin the skin LXe, which produce intense scintillation sig-nals, could often be detected and identified. A rejection2 igure 1: Cross-section of the proposed TPC for nEXO. e ffi ciency of 40% was used for Bi decays in the skinLXe and on surfaces using the time-correlated Po α (the “Bi-Po”) decay pair (half-life of 164 µ s [9]). Sec-ond, it was assumed that γ -ray and β interactions in theskin LXe would neither produce an exploitable signal,nor confuse coincident signals from interactions insidethe TPC.This work reevaluates the above assumptions usingthe open-field-cage nature of nEXO TPC for event re-construction and background discrimination. First, theanalysis quantifies the tagging e ffi ciency for Bi-Po de-cays in the skin LXe. Second, simultaneous scintillationlight in the skin LXe and inside the TPC is modeled indetail, showing that this light can be exploited to rejectcoincident skin γ -ray interactions based on their multi-sited nature. In both cases, the resulting new back-ground estimates for nEXO are compared to those ob-tained with the assumptions from [3].
2. Methodology
The impact of the open-field-cage design was stud-ied using a G eant
Figure 2: Cutaway sketch of the anode section of the TPC. The con-figuration of the charge collection tiles is shown below a cutaway ofthe anode backbone. [4]
The simulated TPC consists of 4896 kg of homoge-neous LXe, isotopically enriched to 90% in the
Xenuclide. The field gradient is controlled by the volt-age di ff erence between the anode and the cathode, and asystem of field-shaping rings assures parallel field linesto enhance charge collection. These components areshown in Fig. 2. The TPC bottom is an opaque cath-ode at high potential, and the top is an anode at groundpotential.Charge is collected by an array of fused silica “tiles”covered electrodes. Fig. 2 shows the placement of thetiles on the underside of the anode surface. All ioniza-tion produced by interactions in the LXe inside the fieldrings is collected on the tiles, no ionization is collectedfrom outside the field rings.The scintillation light is collected by an array ofSiPMs sensitive to the 175 nm wavelength xenon scin-tillation light [11, 12]. The SiPMs are arranged in 24staves outside the field rings, as shown in Fig. 2, for a to-tal photosensitive surface area of ∼ . The top of thestaves are inline with the anode, and the bottoms extend6 cm below the cathode. The SiPMs collect scintillationlight from particle interactions both inside and outsideof the TPC field cage. The nontransparent cathode andanode reduce direct line of sight from the photosensorsto light produced above the anode or below the cathode.3 .2. Event generation Rn, and its radioactive daughters, are well-knownliquid xenon contaminants that have been studied in de-tail [13, 14]. Of particular concern for nEXO is the
Rn-daughter,
Bi, whose decay includes a γ raywith an energy of 2447 . Xe 0 νββ
Q-value ( Q ββ = . ± .
31 keV) [15, 16]. The Bi β decay is followed by a Po α decay with a half-life of164 µ s [9]. By tagging the α decay, this time-correlatedBi-Po decay chain can be identified, and the resultingbackground events rejected.Bi-Po decay chains were simulated both as di ff usedsources in the LXe volume as well as localized sourceson certain surfaces. The simulation was performed inthis way because both neutral and charged daughtersare present in the Rn decay chain [14]. The neu-tral daughters remain in the bulk LXe volume while thecharged drift in the electric field and plate onto nega-tively charged surfaces. The ions in the barrel of theskin LXe plate onto the outer radial surface of the fieldrings, and those inside the TPC field cage plate on thetop surface of the cathode. The charged daughters inthe skin LXe above the anode are assumed to remainin the bulk LXe because the electric field in that regionis zero. Under the cathode, any ions plate onto the un-derside surface of the cathode. The number of
Bidecays from each region was calculated as in [14] us-ing measured ion fractions, mobility, and drift time withthe expected electric fields in LXe for nEXO. This re-sulted in 15.6% of the
Bi decays taking place in theTPC LXe, 5.4% in the skin LXe, 58.9% on the cathodesurface, and 20.1% on the field rings outer surface. Asteady state population of 600
Rn atoms in the totalLXe volume of nEXO was assumed [4].A uniform spatial distribution was assumed for
Bidecays from
Rn daughters that remain in the liquidxenon. Ionized
Bi daughters were simulated on thenearest corresponding surface, at zero depth in the ma-terial. The subsequent
Po nuclei were allowed to re-main on the surface or move into the LXe according toG eant γ -induced backgrounds, inter-actions from the decays of U and
Th and theirdaughters were simulated. These radionuclides con-stitute the dominant background in nEXO. They arepresent as contamination in detector materials, with thelargest contribution from the copper that makes up theTPC vessel. As in [3], the radionuclides simulated forthe
U and
Th decay chains were selected based on
Event Type Location
U Chain TPC Vessel 10
U Chain Internal TPC Components 10
Th Chain TPC Vessel 10
Th Chain Internal TPC Components 10
Bi -
Po Skin LXe Volume 10
Bi -
Po Outer Field Ring Surface 10 Table 1: Event type, location, and number of simulated primary de-cays for this study. the emitted γ -radiation energy >
100 keV and intensity > U and
Th decaychain equilibrium.Using G eant
4, each of the selected decay types weresimulated in su ffi cient quantity to obtain statisticallysignificant numbers of events in the inner 2000 kg LXevolume. The decay type, location, and quantity simu-lated are listed in Table 1. The output of the G eant ff ects. Simulated eventsare analyzed to extract the event parameters of inter-est: multiplicity (Single-Site (SS) or Multi-Site (MS)),distance from the nearest TPC surface (Stando ff dis-tance), and reconstructed event energy. This analysislargely follows the procedure detailed in Ref. [3], al-though some key di ff erences will be called out in thefollowing summary.The multiplicity of the simulated events is determinedfirst. An algorithm groups energy deposits within 3 mmof each other into clusters. The 3 mm cluster size is cho-sen to emulate the discrimination ability projected fornEXO using a separate detailed charge transport simu-lation [17]. To be labeled SS, an event can only haveone reconstructed cluster inside the TPC LXe. This re-quirement separates the 2 νββ and 0 νββ events (whichare primarily SS) from other types of interactions, likeCompton scatters, which are typically MS. Any eventthat has more than one reconstructed cluster is labeledas MS. For this study, SS events with one or more ad-ditional energy deposits in the skin LXe are further la-beled as MS-skin events. This MC-truth value is storedto later assess the e ffi ciency of the analysis.Once all the energy deposits are identified andrecorded the production of scintillation and ionizationquanta for each are explicitly calculated. NEST version2.1 [18, 19] is used to compute the correlated light and4 igure 3: Schematic map of the estimated light collection e ffi ciencyapplied to simulations as a function of position (Z vs Radius). Theseestimated e ffi ciencies are uniform across each region, an approxima-tion based on [3, 4]. The TPC LXe is set at 7%, the barrel of the skinLXe at 10%, and the regions above the anode and below the cathodeat 0.1%. charge quanta based on the electric field at a given lo-cation, and the deposited energy. The electric field isestimated using a cylindrically symmetric electrostaticmodel developed in COMSOL [20].The charge signal is lost in the skin LXe and the lightcollection e ffi ciency varies depending on position, pri-marily for geometrical reasons. The estimated light col-lection e ffi ciencies are uniform across each region, anapproximation based on previous work [3, 4]. The TPCLXe is set at 7%, the barrel of the skin LXe at 10%,and the regions above the anode and below the cathodeat 0.1%. A schematic representation is given in Fig 3.Charge and light collection e ffi ciencies are applied tothe number of scintillation and ionization quanta fromNEST.After all the energy deposits in an event are dividedinto collected light and charge, the event energy is cal-culated following the technique in [21] where the anti-correlation between light and charge signals in LXe [5]is leveraged to generate a rotated energy axis with sig-nificantly better resolution than the individual channels.The e ff ects of instrumental noise in the light and chargesignals are added during this analysis, with noise valueschosen to match the nEXO expected energy resolutionat Q ββ of 1% [4].In Ref. [3], the full description of the complex experi- Figure 4: Position of time correlated Bi-Po interactions (Z vs Radius)produced in the skin LXe. Reconstructed SS events in the TPC LXefrom
Bi coincident γ rays are shown in the TPC LXe (grey). Theposition of coincident α decays from Po in the skin LXe are shownin the skin LXe (blue). The SS event in the TPC LXe is backgroundunless the time correlated α decay is identified. mental backgrounds was distilled down into a single ref-erence value – namely the number of background eventsreconstructed within the FWHM window around Q ββ inthe inner 2000 kg LXe volume. The same simplifyingmetric is used in this analysis. Bi -
Po decays in the skin LXe
There are four regions that contribute
Bi-
Pobackgrounds: the TPC LXe, the skin LXe, the outer sur-face of the field rings, and the top surface of the cathode.Backgrounds from the
Bi decays can be vetoed andrejected if the Po α decay is detectable. Fig. 4 showsthe positions of both the Bi-Po decays and SS energydeposits inside the TPC LXe produced by γ rays from Bi β decays in the skin LXe.Interactions of α particles inside the TPC LXe areeasily distinguished from β and γ -ray interactions basedon the ratio of charge and light signals which di ff er bymore than a factor of 20 [14]. In the skin LXe region,where charge signal is absent, the light signal alone canbe exploited to identify α particle interactions. The dis-tribution of total collected photons from Po α decaysin nEXO’s skin LXe is shown in Fig. 5. Most α decaysresult in a total light detection that is more than 10 timeslarger than that of a 0 νββ event. The exceptions to this5 igure 5: Number of photons detected from Po α decays in the skinLXe (dotted) and on the field ring surfaces (solid). The dot-dashedline marks the maximum light detected from a γ ray with energy at Q ββ in the barrel of the skin LXe. include α decays in the regions with low light collectione ffi ciency above the anode and below the cathode, and α s that lose some or all of their energy inside a structuralcomponent (e.g. in the copper of the field cage rings).If the Po α decay can be tagged in the skin LXe,then a veto can be applied to reject potential Bi time-correlated backgrounds. With an α selection cut on thetotal scintillation light of > × photons, the α tag-ging e ffi ciency is 98% when the α starts in the barrelof the skin LXe. Alpha particles that lose part or all oftheir energy in the detector structure are responsible forthe 2% lost e ffi ciency. Above the anode and below thecathode the tagging e ffi ciency drops to <
1% due to thepoor light collection e ffi ciency. 49% of the α s that de-cay from the field ring surface are successfully tagged.Half of them are contained inside the ring material. MS-skin events have a single energy deposit in theTPC LXe with one or more coincident energy deposi-tions in the skin LXe. These events are typically dueto Compton scattering or emission of multiple coinci-dent γ rays from a single radioactive decay. If the sepa-rate skin interaction is not tagged, these events will con-tribute to the background.Coincident energy deposits in the skin LXe and TPCLXe are not resolved by the photosensor timing. Thesingle scintillation signal and single charge cluster re- Figure 6: Normalized C / L distribution for SS events in the LXe inner2000 kg from decays originating from
Bi (blue) and
Tl (green)contamination in the bulk of the copper of the TPC vessel. The dis-tribution for 0 νββ events is also shown (black). The C / L cut in thisanalysis is shown for reference. sembles a true SS event such as a 0 νββ . However, MS-skin events typically have a lower ratio of the chargesignal amplitude to the light signal amplitude (“C / L” ra-tio) than true SS events. This e ff ect is visible in Fig. 6which shows the distribution of C / L ratios for simu-lated decays originating in the copper TPC vessel. Inthe absence of skin interaction, as in the case of 0 νββ events, an almost-Gaussian distribution of C / L values isobserved, the width of which is determined by the re-combination fluctuations between charge and light pro-duction, the individual resolution for each channel, andNEST’s parametrization of the energy-dependence ofC / L. In contrast, for γ -ray events which can includeskin interactions, a broader distribution of C / L valuesis seen. A considerable fraction of events, 56% of
Biand 69% of
Tl, were MS-skin events and so have re-duced C / L values. For clarity, only the distributions ofdecays from
Bi (
U chain) and
Tl (
Th chain)have been shown since these are the only contributors,48% and 4% respectively, to the 0 νββ background near Q ββ .The e ff ect of the additional light from coincident in-teractions in the skin can be better appreciated in thescatter plot of photons vs electrons detected, as shownin Fig. 7 and Fig. 8 for simulated Bi and
Tl de-cays respectively. Only events that have a single recon-structed charge cluster, within the inner 2000 kg LXe,are shown. In closed-cage TPC (left plots), which wasassumed in the previous analysis, all SS events lie along6he same C / L line, with some variation around that linedue to anticorrelated recombination fluctuations. In anopen-cage TPC (right plots), many events occur withdisproportionately more photons. These extra photonsare produced by interactions in the skin LXe.The features in the distribution of events outside ofthe anticorrelation band can be understood by consider-ing the two processes responsible for generating eventswith a single deposit in the TPC LXe in coincidencewith one or more deposits in the skin LXe: Comptonscattering of a single γ ray and coincident γ rays.When a single γ ray Compton scatters in the skin LXevolume followed by a photoelectric interaction in theTPC LXe volume, in an open-cage design, the scintil-lation signal sums to a value dependent on the energyof the initial γ ray, while only the fraction of chargedeposited in the TPC LXe is collected. This givesrise to horizontal band structures, like the one around6000 detected photons in the right plot of Fig. 8, whichcorresponds to the 2615 keV γ ray from Tl decay.Around 87% of
Tl induced background events nearthe Q ββ ± FWHM / ff erent skin regions.Radioactive decays that emit multiple coincident γ -ray emissions are responsible for events like those inthe structure protruding above the spot at ∼ γ ray undergoing pho-toelectric absorption in the TPC LXe, with a coinci-dent γ ray emitted in the same decay (85% of the timethe 583 keV γ ray [22]) interacting in the skin LXe.This type of events represents 13% of Tl decays inthe Q ββ ± FWHM / γ ray from the Bi decaybecause these decays nearly always have a γ -ray multi-plicity of 1.As extreme values of the C / L ratio are produced onlyby backgrounds, events with these values should be re-jected. Here, we assess the impact of a simple cut thatremoves events with C / L <
15, as shown by the redline in Fig. 8 and Fig. 7. The value for this C / L cutwas defined by requiring >
99% 0 νββ signal acceptancerate. Experimentally, this value would be determinedfrom calibration and 2 νββ data as was done in the re-cent EXO-200 analysis [23].The reconstructed energy (using the optimized lin-ear combination of light and charge) is shown in Fig. 9for simulated
Tl and
Bi decays originating in theTPC vessels. Four spectra are overlaid in these plots.Each spectrum shows events with a single reconstructed charge cluster in the inner 2000 kg TPC LXe. The firstspectrum shows the energy that would be obtained in aclosed-cage TPC. The second spectrum shows the en-ergy observed by an open-cage TPC. In this spectrum,due to coincident skin light the energy can be misre-constructed, giving rise to features like the hump on theright of the
Tl peak. The third spectrum shows theopen-cage TPC results after applying the C / L >
15 re-quirement. Finally, the fourth spectrum applies a cutusing MC truth, excluding all events with a skin compo-nent. This shows the hypothetical ideal selection whichremoves all skin-interacting backgrounds. In the regionof interest around Q ββ , no significant di ff erence is ob-served between the di ff erent analyses for Bi decays.On the other hand, MS-skin events account for a largefraction of the background originating from
Tl. Thisis due to shallow Compton scattering in the skin LXeof the 2615 keV γ -ray, and to the coincident 583 keV γ -ray depositing energy in the skin LXe. The addi-tion of coincident skin light to the energy reconstructionimpacts the background rate across the entire detector,not only in the inner 2000 kg. Therefore, the relativeimpact on the expected background contribution in the Q ββ ± FWHM / Tl, a non-linear increase in the SS backgroundsnear the Q ββ was observed when considering interac-tions in the TPC LXe volume closer to the edges of thefield cage. This behavior is understood by looking ata breakdown by event type of these interactions. Thebackground in the inner 2000 kg of LXe is made upof mostly (79%) events where the with primary γ rayCompton scatters in the skin LXe, then deposits the restof its energy in the TPC LXe. The remainder of eventsare divided between decays with multiple γ rays (13%),where the primary γ ray (2615 keV) deposits energy in-side the TPC LXe and a second coincidence γ ray de-posits energy in the skin LXe, and events where the pri-mary γ ray Compton scatters in the TPC LXe, followedby photoelectric adsorption in the skin LXe (8%). TheTPC LXe volume outside of the inner 2000 kg has a dif-ferent profile of event types. It has a larger (49%) contri-bution from those events where the primary γ ray scat-ters first in the TPC LXe then the skin LXe. They arepredominately events where the energy deposited in theTPC is near the Compton edge for the 2615 keV γ ray,located at 2382 keV. By itself, a Compton scatter at thisenergy is too far away from Q ββ to be included in thebackground count. However, when including additionallight from the subsequent interaction in the skin, theseevents are misreconstructed near the Q ββ value. Becausethe energy deposited in the skin LXe is typically small7 igure 7: Detected Photons vs Detected Electrons for Bi decays inside the copper of the TPC vessel. Simulations assumed closed-cage TPC(left) and open-cage TPC (right) in the inner 2000 kg LXe. The Q ββ rotated energy value and C / L cut are shown for reference.Figure 8: Detected Photons vs Detected Electrons for
Tl decays inside the copper of the TPC vessel. Simulations assumed closed-cage TPC(left) and open-cage TPC (right) in the inner 2000 kg LXe. The value for Q ββ in the rotated energy axis and the C / L cut are shown for reference. igure 9: Reconstructed energy of SS events in the inner 2000 kg TPC LXe for Tl (left) and
Bi (right) decays inside the copper of the TPCvessel, under di ff erent conditions. The distribution obtained for a closed-cage TPC (green), is compared to that obtained in the open-cage TPC withand without removing events with C / L >
15 (black and blue lines respectively). The results from the ideal case where interactions in the skin LXecould be perfectly reconstructed and identified are also shown (magenta). The insets show a zoom of the distributions near the Xe Q ββ . ( ∼ ffi cient to bringthe C / L value below the cut threshold. Thankfully, theseevents predominantly appear in the outer volume of theTPC LXe and have a limited impact on the sensitivity.In the simulation performed for this work, no angu-lar correlation was assumed between coincident γ -raysfrom decays such as Tl [24]. This e ff ect was evalu-ated in the context of EXO-200 calibration and foundinsignificant [25].
3. Results and Discussion
We now turn our attention to the impact that the open-cage design has on the background in nEXO. As dis-cussed earlier, the nEXO 0 νββ sensitivity estimationspreviously assumed that for interactions in the skin,only α particles could be detected. All signals from γ -ray interactions in the skin were neglected. In do-ing so, energy misreconstructions were not considerednor were possible background reductions from exploit-ing charge-to-light features.The background resulting from the prior simplifiedanalysis is summarized in the first column of Table 2.This summary includes only the main background com-ponents in nEXO relevant to the analysis of interac-tions in the skin LXe. These include interactions fromthe U and
Th decay chains from bulk contami-nation of the materials in the TPC vessel and internalcomponents. Contributions from daughters of dissolved
Rn are also included. Values are normalized to thetotal background budget. Backgrounds were estimated near Q ββ for the central region of the detector (inner2000 kg), since this region dominates the detector sen-sitivity [3].The second column of Table 2 shows the e ff ect onthe background budget of exploiting the light collectedfrom interactions in the skin LXe in an open-cage TPC,calculated using the method discussed in the previoussection. The C / L <
15 cut was applied to reject interac-tions with a skin component while retaining more than99% of 0 νββ events. For identification of Bi-Po eventsin the skin, the analysis considered a veto of one secondbefore any α decay in the skin. The veto time is signif-icantly longer than the Po half-life of 163 µ s, and sothe Bi-Po tagging e ffi ciency is dominated by the abilityto observe and tag the α decay in the skin. Even whenan α is not visible in the skin, sometimes the light signalfrom the coincident β particle emitted during the Bidecay is su ffi cient for rejection via the C / L cut. Theanalysis resulted in a Bi-Po rejection e ffi ciency of 76%and 55% for events originating in the skin LXe and fieldring surfaces, respectively.These results confirm that the background estimatedunder the skin-interaction assumptions in the prior studywere slightly conservative when compared to the re-sults from this new analysis that fully exploits the open-cage design of nEXO. The previous assumption that co-incident γ -ray interactions in the skin would not im-pact the total backgrounds proved largely correct in thismore detailed analysis. This minimal impact stems fromthe small fraction of Bi decays that have multi-sitedinteractions involving the skin and the e ffi cacy of the9 / L ratio cut at removing
Tl skin-interacting back-grounds. This study’s detailed accounting of skin Bi-Poevents shows the background from
Rn dissolved inLXe can be rejected more e ffi ciently than assumed inthe simplified analysis.Finally, we consider the ideal case in which the detec-tor can identify with 100% e ffi ciency all MS-skin eventsand Po α emission in the skin. This is shown in thelast column of Table 2. In this case, the Th compo-nent is more than halved, while the
U component islargely unchanged. The decrease in the
Th compo-nent is because of shallow Compton scatter in the skinof the 2615 keV γ ray and decays with γ -ray multiplic-ity larger than one. These are MS-skin whose skin in-teractions are too low in energy to be tagged under theassumptions in the non-ideal, open-field-cage analysis,but which are caught in this ideal scenario. The addi-tional improvement in the Rn background is primar-ily due to rejection of Bi-Po decays in the LXe regions(above the anode, below the cathode) where the lightcollection e ffi ciency is small ( ∼
4. Conclusions
This study has established that an optically-open fieldcage design introduces some complexity in the event re-construction of a liquid xenon TPC but can be used toidentify and reject skin-interacting backgrounds. Time-correlated Bi-Po backgrounds can be removed by identi-fying light signals from the
Po decays in the skin andon nearby surfaces. Coincident backgrounds are not sig-nificantly di ff erent between open-cage and close-cagedesigns. As Bi backgrounds rarely interact in the skinand
Tl skin-interacting backgrounds can be rejectedusing a simple cut. Together, these result in a 5% re-duction in background over the estimate in [3]. This re-duction does not lead to an appreciable improvement innEXO’s sensitivity to 0 νββ
Xe half-life, as nEXO’shalf-life sensitivity ( T ν / ) only scales with the back-ground rate in the inner 2000 kg as ( B ) as T ν / ∝ B − . .Continued improvements in the ability to reconstructand tag the light in the skin LXe could provide addi-tional background reduction, up to 12% in the idealcase of perfect skin tagging. Better understanding ofevent identification strengthens nEXO’s ability to as-sess and control its backgrounds. Furthermore, the abil-ity to better identify skin LXe interactions may becomeimportant during calibration measurements due to pile-up. nEXO plans to calibrate the detector spatial light-collection e ffi ciency using a set of intense γ -ray sources placed outside of the TPC vessel. In order to reach asu ffi cient count rate of “deep” events in the inner regionof the LXe volume, an interaction rate of up to 1.6 kHzis expected in the detector. The majority of those inter-actions happen in the outer regions of xenon, includingthe skin LXe. In this situation, rejection of interactionwith a skin component can help in isolating the deepevents necessary for the calibration. One approach thatcan be tried consists of exploiting the spatial distributionof collected photons on the SiPMs. The highly concen-trated pattern from a skin LXe γ ray may provide a rec-ognizable feature when compared to the di ff use patternfrom a centralized deep event. Acknowledgements
Support for nEXO comes from the the O ffi ce of Nu-clear Physics of the Department of Energy and NSF inthe United States, from NSERC, CFI, FRQNT, NRC,and the McDonald Institute (CFREF) in Canada, fromIBS in Korea, from RFBR (18-02-00550) in Russia, andfrom CAS and NSFC in China. LLNL-JRNL-814563 References [1] C. Patrignani, et al. (Particle Data Group), Review of Parti-cle Physics, Chin. Phys. C40 (2016) 100001. doi: .[2] E. Majorana, Theory of the symmetry of electrons andpositrons, Nuovo Cimento 14 (1937) 171–184. doi: .[3] J. B. Albert, et al. (nEXO), Sensitivity and Discov-ery Potential of nEXO to Neutrinoless Double Beta Decay,Phys. Rev. C97 (2018) 065503. doi: . arXiv:1710.05075 .[4] S. A. Kharusi, et al. (nEXO), nEXO Pre-Conceptual DesignReport (2018). arXiv:1805.11142 .[5] E. Conti, et al. (EXO-200), Correlated fluctuations betweenluminescence and ionization in liquid xenon, Phys. Rev.B68 (2003) 054201. doi: . arXiv:hep-ex/0303008 .[6] E. Aprile, et al. (XENON), The XENON1T Dark Matter Ex-periment, Eur. Phys. J. C77 (2017) 881. doi: . arXiv:1708.07051 .[7] B. J. Mount, et al., LUX-ZEPLIN (LZ) Technical Design Report(2017). arXiv:1703.09144 .[8] M. Auger, et al. (EXO-200), The EXO-200 detector, part I: De-tector design and construction, JINST 7 (2012) P05010. doi: . arXiv:1202.2192 .[9] M. Shamsuzzoha Basunia, Nuclear Data Sheets for A = .[10] S. Agostinelli, et al. (GEANT4), GEANT4: A Simulationtoolkit, Nucl. Instrum. Meth. A506 (2003) 250–303. doi: .[11] A. Jamil, et al. (nEXO), VUV-sensitive Silicon Photomultipli-ers for Xenon Scintillation Light Detection in nEXO, IEEETrans. Nucl. Sci. 65 (2018) 2823–2833. doi: . arXiv:1806.02220 . ackground FractionSimplified Detailed Open-cage Perfect SkinAnalysis [3] TPC Analysis LXe Tagging U TPC Components 0 .
48 0 .
45 0 . Th TPC Components 0 .
039 0 .
039 0 . Rn Skin LXe 0 .
043 0 . . Rn Field Ring Surfaces 0 .
026 0 . Table 2: Background contributions from SS events to the Q ββ ± FWHM / ff ects of an open-cage design. Results in the lastcolumn are obtained under the ideal case where interactions in the skin LXe could be perfectly reconstructed and identified.[12] G. Gallina, et al. (nEXO), Characterization of the Hama-matsu VUV4 MPPCs for nEXO, Nucl. Instrum. Meth. A940 (2019) 371 – 379. doi: . arXiv:1903.03663v3 .[13] J. B. Albert, et al. (EXO-200), Investigation of radioactivity-induced backgrounds in EXO-200, Phys. Rev. C92(2015) 015503. doi: . arXiv:1503.06241 .[14] J. B. Albert, et al. (EXO-200), Measurements of the ion fractionand mobility of α - and β -decay products in liquid xenon usingthe EXO-200 detector, Phys. Rev. C92 (2015) 045504. doi: . arXiv:1506.00317 .[15] M. Redshaw, E. Wingfield, J. McDaniel, E. G. Myers, Massand double-beta-decay Q value of Xe-136, Phys. Rev. Lett. 98(2007) 053003. doi: .[16] P. M. McCowan, R. C. Barber, Q-value for the double-beta de-cay of Xe-136, Phys. Rev. C82 (2010) 024603. doi: .[17] Z. Li, et al. (nEXO), Simulation of charge readoutwith segmented tiles in nEXO, J. Instrum. 14 (2019)P09020–P09020. doi: . arXiv:1907.07512v2 .[18] M. Szydagis, J. Balajthy, J. Brodsky, J. Cutter, J. Huang, E. Ko-zlova, B. Lenardo, A. Manalaysay, D. McKinsey, M. Mooney,J. Mueller, K. Ni, G. Rischbieter, M. Tripathi, C. Tunnell, V. Ve-lan, Z. Zhao (NEST), Noble Element Simulation Techniquev2.0, 2018. doi: .[19] J. Brodsky, C. Tunnell, M. Szydagis, J. Balajthy, V. Velan,J. Huang (NEST), NESTCollaboration: Geant4 IntegrationFixes and Updates, 2019. doi: .[20] COMSOL c (cid:13) , , 2019.[21] J. B. Albert, et al. (EXO-200), Improved measurementof the 2 νββ half-life of Xe with the EXO-200 detector,Phys. Rev. C89 (2014) 015502. doi: . arXiv:1306.6106 .[22] M. J. Martin, Nuclear Data Sheets for A = .[23] G. Anton, et al. (EXO-200), Search for neutrinoless double- β decay with the complete EXO-200 dataset, Phys. Rev. Lett. 123(2019) 161802. doi: .[24] L. C. Biedenharn, M. E. Rose, Theory of angular correlation ofnuclear radiations, Rev. Mod. Phys. 25 (1953) 729–777. doi: .[25] R. MacLellan, A. Piepke, D. Auty, On the angular correlationbetween γ -rays, EXO-200 Internal Note, 2012.-rays, EXO-200 Internal Note, 2012.